Last updated: September 2026
Format guide for engineers building on Anthropic Claude — Messages API, tool use, MCP, Claude on Bedrock/Vertex, and production agent systems. Distinct from generic "LLM engineer" resumes that only list OpenAI.
Use reverse-chronological, single-column layout. Each bullet pairs Claude-specific stack with a metric: latency, token cost, task success rate, or human eval score. Keywords hiring teams scan for: Anthropic API, Claude 3.5/4 Sonnet & Opus, tool use, structured outputs, MCP, prompt caching, batch API, Bedrock/Vertex deployment, guardrails, and eval regression suites.
Related: AI professionals hub · Agentic AI engineer resume · Senior LLM engineer resume · ChatGPT vs resume builder
Example bullet: "Shipped Claude Sonnet agent with 8 MCP tools for internal IT support, resolving 34% of tickets without human escalation while keeping p95 latency under 2.1s via prompt caching."
Claude API · Anthropic SDK · Tool use & structured outputs · MCP · LangChain/LangGraph with Claude · Bedrock & Vertex · Python/TypeScript · RAG grounding · Prompt systems · Token/cost optimization · SOC 2–aware logging
Anthropic API, tool use, MCP, deployment target (Bedrock/Vertex), eval metrics, and cost/latency outcomes — not just "used Claude."
Claude roles want Anthropic-specific patterns (caching, MCP, extended context). Mirror the JD; list multi-model experience only if true.
Yes — it's a high-signal keyword for Claude/agent hiring in 2026 when you have hands-on MCP server or client work.